FAT* is an international and interdisciplinary peer-reviewed conference that seeks to publish and present work examining the fairness, accountability, and transparency of algorithmic systems. The FAT* conference solicits work from a wide variety of disciplines, including computer science, statistics, the humanities, and law. FAT* welcomes submissions that touch on any of the following topics (broadly construed):
The proceedings for FAT* 2018 will appear in the Proceedings of Machine Learning Research.
To ensure that all submissions to FAT* are reviewed by a knowledgable and appropriate set of reviewers, the conference is divided into topics. Authors must choose one or more of the following topics when they register their submissions:
FAT* 2018 offers authors the choice of archival and non-archival paper submissions. Archival papers will appear in the published proceedings of the conference, if they are accepted; conversely, accepted non-archival papers will only appear as abstracts in the proceedings. FAT* offers a non-archival option to avoid precluding the future submission of these papers to area-specific journals. Note that all submissions will be judged by the same quality standards, regardless of whether the authors choose the archival or non-archival option. Furthermore, reviewers will not be told whether submissions under review are archival or not, to avoid influencing their evaluations.
Authors of all accepted papers must present their work at the FAT* 2018 conference, regardless of whether their paper is archival or non-archival.
FAT* 2018 is using HotCRP as the paper submission and reviewing system. The submission website is available here.
Submitted papers must be 8-10 pages (including all figures, tables, and references), plus unlimited pages for appendices. Reviewers will not be required to read material in appendices, so authors are encouraged to use them judiciously.
Papers should be formatted using the 2017 ACM Master Article Template. For LaTeX users, choose
format=sigconf. This is the typical, two-column proceedings-style template. Authors do not need to include terms, keywords, or other front matter in their submissions. ACM also makes a Word template available; however, authors who wish to eschew the ACM Word template may submit manuscripts in two-column format, with one inch margins, 9 point Times New Roman font. Note that all accepted papers will need to be reformatted to the camera-ready template, which will be announced later.
Authors will select (1) one or more topics for their submission and (2) whether their submission is archival or non-archival during paper registration. The selected topic(s) will determine the pool of PC members who will review the submission. The PC Chairs reserve the right to move submissions between topics if the PC feels that a submission has been misclassified.
FAT* uses a double blind review process. Authors must omit their names and affiliations from submissions, and avoid obvious identifying statements. Citations to the authors' own prior work should be made in the third-person. Submissions that do not comply with this policy will be rejected without review.
Confidentiality of submitted material will be maintained. Upon acceptance, the titles, authorship, and abstracts of papers will be released prior to the conference.
FAT* welcomes submission of work that has previously appeared in non-archival venues. These works may be submitted as-is or in an extended form. FAT* also welcomes full paper submissions that extend previously published short papers (e.g. from workshops). Authors must still take care to comply with the double blind reviewing requirements when submitting extensions of prior work.
Submitting authors make the following representations about their work (drawn from the ACM Author Policy Representations):
Papers that (1) describe experiments with users and/or deployed systems (e.g., websites or apps), or that (2) rely on sensitive user data (e.g., social network information), must follow basic precepts of ethical research and subscribe to community norms. These include: respect for privacy, secure storage of sensitive data, voluntary and informed consent if users are placed at risk, avoiding deceptive practices when not essential, beneficence (maximizing the benefits to an individual or to society while minimizing harm to the individual), risk mitigation, and post-hoc disclosure of audits. When appropriate, authors are encouraged to include a subsection describing these issues. Authors may want to consult the Menlo Report for further information on ethical principles, the Allman/Paxson IMC'07 paper for guidance on ethical data sharing, and the Sandvig et al. '14 paper on the ethics of algorithm audits.
Note that submitting research for approval by each author’s institutional ethics review body (IRB) may be necessary in some cases, but by itself may not be sufficient. In cases where the PC has concerns about the ethics of the work in a submission, the PC will consider the ethical soundness and justification of the submission, just as it does its technical soundness. The PC takes a broad view of what constitutes an ethical concern, and authors agree to be available at any time during the review process to rapidly respond to queries from the PC Chairs regarding ethical considerations. Authors unsure about topical fit or ethical issues are welcome to contact the PC Chairs.
This CfP borrows liberally from the FAT/ML 2017, FATWEB 2017, IMC 2017, Ethics in NLP 2017, and CSCW 2018 CfPs. The PC Chairs extend our thanks to the organizers of these workshops and conferences.